Jin Lei is from Oufei Temple
Qubits | Official account QbitAI
Just now, this year's ACM SIGSOFT Outstanding Doctroal Dissertation Award has been released.
This academic award, which has only one place a year, was won by the Chinese Wing Lam (Lin Yongzheng).
During his PhD at UIUC (University of Illinois at Urbana-Champaign), he was co-supervised by Professor Tao Xie (now Professor of Study at Peking University) and Professor Darko Marinov.
According to the official introduction of ACM, Lin Yongzheng won the award this time because of his outstanding contributions to software engineering:
Professor Xie Tao also sent his blessings on this:
Award-winning papers
The research behind Lin Yongzheng winning this "award" is his graduation thesis when he studied for his Ph.D. at UIUC , Detecting Characterizing, And Taming Flaky Tests.
The background of this research is that with the development of technology, regression testing has gradually become an important part of software testing.
For example, every code submission, software integration, and product delivery requires regression testing to verify.
It can be said that regression testing is a kind of "touchstone" in these processes.
But it is conceivable that when the product function increases, the number of regression test cases will also increase.
Therefore, automating regression testing is a trend towards greater efficiency.
In the process, however, Flaky Tests has become another difficult problem.
(Flaky Tests refers to tests that are sometimes successful and sometimes unsuccessful, and are more unstable when both the test object and the test conditions are unchanged.) ）
Because of the Flakiness nature inherent in automated regression testing, regression testing can be difficult to achieve 100% stability, and the more frequently the use case is executed, the more this instability will be amplified.
Lin Yongzheng's paper revolves around Flaky Tests, which mainly does three aspects of work:
First, a new technique for detecting Flaky Tests is proposed, which allows developers to pre-prevent Flaky Tests from influencing the results of regression tests.
Second, new technologies to describe Flaky Tests are proposed to help developers better understand their Flaky Tests.
Finally, the question of new technologies to tame Flaky Tests is raised so that Flaky Tests does not mislead developers in regression testing by adapting to flakiness.
In terms of detecting Flaky Tests, Lin's work proposes a framework called iDFlakies.
It can be used to detect and classify Flaky Tests locally:
Use cases for running iDFlakies
Using this framework, Lin yongzheng applied it in 683 projects; in addition, he provided a dataset containing 422 Flaky Tests for research.
According to the dataset, 50.5% of Flaky Tests are order-dependent (OD), while 49.5% are non-deterministic (NOD).
Moreover, Lin Yongzheng's research also found that running a random class method configuration can detect the most unstable (flaky) tests overall.
In describing Flaky Tests, the paper pioneered automated tools to help developers debug Flaky Tests failures.
In this step, the paper also examines the effect of test order on NOD testing: even the same order can determine the success or failure of the test.
Finally, in taming Flaky Tests, Lin's research was the first to propose that automation techniques could reduce the number of spurious failures for OD tests by 73 percent.
At the same time, the first automation technology was proposed to speed up the AW (async-waiting) type Flaky Tests (38 percent faster).
For details of the paper, see the link at the end of the article.
Who is Lin Yongzheng?
Wing Lam was born in Hong Kong and grew up in Seattle, USA.
He studied at the University of Washington in Seattle as an undergraduate, and was recruited by Professor Xie Tao to study in the UIUC Research Group for a phD.
Coincidentally, Lin Yongzheng's supervisor when he was doing scientific research during his undergraduate period was Professor Xie Tao's doctoral supervisor David Notkin.
Lin Yongzheng was initially interested in the direction related to Android testing, so he had contact with Professor Xie Tao very early.
Later, because of the topic of his doctoral dissertation, he intersected with Professor Drako Marinov, Professor Xie Tao's colleague at UIUC, and finally decided to jointly supervise him.
It is not difficult to see from Lin Yongzheng's doctoral dissertation that Professor Xie Tao has a profound influence on his scientific research. Lin Yongzheng recalled:
I clearly remember that in the first few years of my PhD, Professor Xie Tao spent a long time with me, and from this time I learned how to shape myself and let me grow into the scientific researcher I am now.
For example, Professor Xie Tao once said to me that I should not just "cook" my research, but let him "smell" or "taste" the soup.
And Lin Yongzheng also said that Professor Xie Tao often encouraged him to think more about the overall situation of his work and have a bigger dream about his work:
I hope to instill such an ideal in my future students.
Today, Lin Yongzheng is an assistant professor at George Mason University in the United States.
About the ACM SIGSOFT Distinguished Doctoral Dissertation Award
The ACM SIGSOFT Distinguished Doctoral Dissertation Award is awarded annually to the author of a distinguished doctoral dissertation in the field of software engineering.
The authors of the winning papers will be invited to present abstracts in Software Engineering Notes (SEN).
The award, which includes a $1,000 honorarium and certificate, is presented at ICSE (International Conference on Software Engineering), the premier conference in software engineering.
Address of thesis:
https://www.ideals.illinois.edu/handle/2142/113017
Reference Links:
[1]https://www.sigsoft.org/awards/dissertationAward.html
[2]https://cs.gmu.edu/~winglam/
[3]http://www.51testing.com/html/45/n-4462645.html